120 Rhodora [Jun 
theory of segregation as would at first seem to be the case. According 
to the Mendelian theory, all unit characters tend to segregate with 
successive generations produced from a hybrid, and to return to the 
pure parent form, but this does not mean that this ultimately pure 
character will necessarily be combined with another ultimately pure 
character from the same parent. In many cases there will be a combi- 
nation of an ultimately constant character of one parent with a simi- 
larly constant character of the other parent so that all the resulting 
offspring in succeeding generations will be alike and will possess both 
of these characters. As the ordinary expression goes, it will be a 
permanent hybrid, and thus a new race will be established. Whether 
we should attempt to distinguish all of these new races by names, or 
simply refer to them as hybrids of *so and so" in which *such and 
such” characters of each parent are represented, is an open question. 
Considering the uncertainty, without culture, as to whether the 
various supposed hybrids breed true or not, it has seemed to the 
writer greatly in the interest of clearness to refrain from giving them 
names. . 
Thus the present treatment will still be disappointing to those who, 
like most of us, would like to be able to refer every specimen to a 
definite category. The writer believes, however, that this may be 
more nearly accomplished under the present treatment than under 
any other method of treating Amelanchier of which he is aware. We 
now have eight well defined categories bearing specific names. Other 
specimens which do not fall into these may be definitely labeled as 
hybrids! between some two of these parents. If the hybrids were 
described as species, a very large number of categories would be 
required to represent the various Mendelian combinations, and thus 
these categories themselves would differ from each other to such a 
slight degree as to render the classification unusable in actual practice. 
Amelanchier, like Rubus, is a group in which it will never be possible 
to have the clearly cut condition found in so many genera where speci- 
mens will fall easily into one or another specific category. 
1 The word hybrid is here used in a sense, common in taxonomy, but differing 
from that adopted by the modern plant breeders. It signifles any individual which 
is the result of crossing and which presents some of the traits of one parent side 
by side with some traits of the other parent. It may or may not breed true. The 
Mendelian students, on the other hand, use the term Aybrid to denote an individual 
of the Dr group, which will not breed true but which will segregate in successive 
generations. 
